INDEX

Explanations

words related to political figures, specifically those with the name "Clinton" and "Biden."

oai_token-act-pair · gpt-3.5-turbo

references to specific individuals, particularly their names and titles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 exha

-0.81

ãħĭ

-0.77

thora

-0.76

Pwr

-0.76

pmwiki

-0.75

rul

-0.71

HAHAHAHA

-0.71

«ĺ

-0.71

 ILCS

-0.70

 murd

-0.70

POSITIVE LOGITS

iden

1.25

ners

0.96

ovo

0.88

vier

0.87

ned

0.85

fold

0.85

unci

0.81

ning

0.80

ception

0.78

ovic

0.77

Activations Density 0.011%

words related to political figures, specifically those with the name "Clinton" and "Biden."

references to specific individuals, particularly their names and titles

No Comments

No Known Activations