INDEX

Explanations

mentions of political figures holding elected positions

oai_token-act-pair · gpt-3.5-turbo

references to political incumbents and their status

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

kie

-0.83

wagen

-0.79

acca

-0.78

atche

-0.78

orgetown

-0.78

gas

-0.75

ombs

-0.74

anguage

-0.74

istic

-0.72

hran

-0.71

POSITIVE LOGITS

 incumbent

0.98

 frontrunner

0.87

 challenger

0.81

 presidents

0.78

 occupant

0.78

 president

0.78

 challengers

0.74

 governor

0.72

 loser

0.70

 mayor

0.69

Activations Density 0.045%

mentions of political figures holding elected positions

references to political incumbents and their status

No Comments

No Known Activations