INDEX

Explanations

the beginning of textual content ("<bos>")

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Configuration

jbloom/Gemma-2b-Residual-Stream-SAEs/gemma_2b_blocks.6.hook_resid_post_16384_anthropic_fast_lr

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

HuggingFaceFW/fineweb

Features

16,384

Data Type

torch.float32

Hook Point

blocks.6.hook_resid_post

Architecture

standard

Context Size

1,024

Dataset

HuggingFaceFW/fineweb

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 massacres

-0.97

corruption

-0.86

 atrocities

-0.86

 bombings

-0.86

 disinformation

-0.85

 insurgents

-0.85

 insurgency

-0.84

ABORT

-0.82

 fascism

-0.81

ायर

-0.80

POSITIVE LOGITS

<bos>

16.55

expandindo

2.53

GEBURTSDATUM

2.48

 betweenstory

2.38

 encomp

2.34

Autoritní

2.33

 intersper

2.31

 Administrativna

2.28

تقاوى

2.27

 dispen

2.26

Activations Density 0.956%

the beginning of textual content ("<bos>")

No Comments

No Known Activations