INDEX

Explanations

terms related to past events and histories

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 10-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_10/width_16k/average_l0_77

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.10.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 purpoſe

-0.82

 myſelf

-0.81

 raiſ

-0.80

 Anſ

-0.77

 pleaſure

-0.75

 auffi

-0.73

 ſche

-0.72

thée

-0.72

 itſelf

-0.71

IntoConstraints

-0.70

POSITIVE LOGITS

 past

0.71

ex

0.64

 former

0.62

Former

0.56

 Past

0.56

 Former

0.56

past

0.55

 archived

0.54

 previous

0.54

 PREVIOUS

0.53

Activations Density 0.558%

terms related to past events and histories

No Comments

No Known Activations