INDEX

Explanations

mentions of political matters or terms

oai_token-act-pair · gpt-3.5-turbo

references to political issues or contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

actory

-0.84

imus

-0.82

imates

-0.82

olen

-0.82

ibles

-0.81

tered

-0.81

upon

-0.77

Customer

-0.74

eret

-0.74

ergic

-0.74

POSITIVE LOGITS

 correctness

1.33

 upheaval

0.98

 unrest

0.92

 affiliation

0.91

 persuasion

0.91

 affili

0.91

 subdivision

0.89

 considerations

0.88

 leaders

0.87

 activism

0.87

Activations Density 0.036%

mentions of political matters or terms

references to political issues or contexts

No Comments

No Known Activations