INDEX

Explanations

phrases related to American politician Nancy Pelosi

oai_token-act-pair · gpt-3.5-turbo

mentions of Nancy Pelosi

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

tered

-0.77

=-=-=-=-

-0.76

Redditor

-0.76

orically

-0.75

upon

-0.74

points

-0.74

=-=-

-0.74

orical

-0.72

orescent

-0.71

ioch

-0.71

POSITIVE LOGITS

 Pelosi

1.24

 Drew

0.88

Lan

0.86

 Reagan

0.83

 Grace

0.79

vier

0.78

GE

0.77

 Claus

0.74

 McMahon

0.72

 Gibbs

0.72

Activations Density 0.024%

phrases related to American politician Nancy Pelosi

mentions of Nancy Pelosi

No Comments

No Known Activations