INDEX

Explanations

This neuron appears to activate with certain combinations of characters or possibly words that are not consistently related across the different snippets, suggesting that it may be picking up on a very specific and potentially meaningless pattern in the text, or several patterns.

oai_token-act-pair · gemini-2.0-flash

New Auto-Interp

Configuration

google/gemma-scope-2b-pt-transcoders/layer_25/width_16k/average_l0_41

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.25.ln2.hook_normalized

Architecture

jumprelu_transcoder

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 '\\;'

-0.60

TintMode

-0.55

writeFieldEnd

-0.52

хьтан

-0.52



-0.50

ExecuteReader

-0.49

 تضيفلها

-0.47

éez

-0.47

expandindo

-0.46

 صوتيه

-0.45

POSITIVE LOGITS

 AssemblyTitle

0.64

intenance

0.59

richTextPanel

0.56

agissait

0.51

 esboço

0.51

uxxxx

0.51

 cherchés

0.50

ungsbedingungen

0.49



0.49

生平

0.48

Activations Density 0.684%

This neuron appears to activate with certain combinations of characters or possibly words that are not consistently related across the different snippets, suggesting that it may be picking up on a very specific and potentially meaningless pattern in the text, or several patterns.

No Comments

No Known Activations

This neuron appears to activate with certain combinations of characters or possibly words that are not consistently related across the different snippets, suggesting that it may be picking up on a very specific and potentially meaningless pattern in the text, or several patterns.

No Comments

No Known Activations