INDEX

Explanations

mentions of names or proper nouns, possibly in the context of news articles

oai_token-act-pair · gpt-3.5-turbo

elements indicating citation or reference to articles and their sources

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 ILCS

-0.83

 Â¥

-0.66

 Slayer

-0.62

MLA

-0.61

 Provincial

-0.61

 licence

-0.61

CFL

-0.59

Â£

-0.59

tein

-0.59

Ont

-0.58

POSITIVE LOGITS

 POLITICO

1.30

Trump

1.07

 Politico

1.07

Democrats

1.01

Hillary

1.01

Senate

0.98

 Priebus

0.98

Republicans

0.98

 Trump

0.96

Clinton

0.96

Activations Density 0.258%

mentions of names or proper nouns, possibly in the context of news articles

elements indicating citation or reference to articles and their sources

No Comments

No Known Activations