INDEX

Explanations

references to political positions or roles, especially those related to opposition or alternative leadership

oai_token-act-pair · gpt-3.5-turbo

references to "shadow" roles or positions in political contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

urses

-0.79

ickr

-0.78

ktop

-0.72

keye

-0.72

artney

-0.72

renheit

-0.70

anchester

-0.69

aii

-0.69

OPLE

-0.68

TAIN

-0.68

POSITIVE LOGITS

boxing

1.01

moon

1.01

loo

0.95

runners

0.88

fax

0.84

flame

0.78

fell

0.77

shadow

0.76

wra

0.76

runner

0.76

Activations Density 0.044%

references to political positions or roles, especially those related to opposition or alternative leadership

references to "shadow" roles or positions in political contexts

No Comments

No Known Activations