INDEX

Explanations

mentions of political positions and titles

oai_token-act-pair · gpt-3.5-turbo

references to "shadow" cabinet members or positions

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

urses

-0.92

ickr

-0.82

awaru

-0.80

olitan

-0.79

ktop

-0.75

orsi

-0.74

renheit

-0.74

anchester

-0.73

erto

-0.71

keye

-0.70

POSITIVE LOGITS

moon

0.93

 shadow

0.84

 Shadow

0.83

shadow

0.82

runners

0.81

boxing

0.79

loo

0.78

dust

0.76

runner

0.75

busters

0.75

Activations Density 0.013%

mentions of political positions and titles

references to "shadow" cabinet members or positions

No Comments

No Known Activations