INDEX

Explanations

political entities and terms related to elections and government positions, especially the United States Senate

oai_token-act-pair · gpt-3.5-turbo

references to political offices and elections

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

oteric

-0.78

 sensit

-0.73

Feed

-0.72

 divul

-0.70

ppelin

-0.70

tip

-0.69

autical

-0.68

llah

-0.68

iger

-0.66

 shelling

-0.66

POSITIVE LOGITS

 Elections

1.02

 elections

0.92

election

0.87

 voting

0.84

 challenger

0.83

 election

0.78

cand

0.78

 incumb

0.78

 challengers

0.77

 Voting

0.77

Activations Density 0.257%

political entities and terms related to elections and government positions, especially the United States Senate

references to political offices and elections

No Comments

No Known Activations