INDEX

Explanations

words related to political figures and news outlets

oai_token-act-pair · gpt-3.5-turbo

mentions of specific individuals associated with political commentary and media

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 ants

-0.63

urable

-0.62

 Superintendent

-0.62

 superintendent

-0.62

 Branch

-0.61

tics

-0.61

 starved

-0.61

 rods

-0.61

urated

-0.60

 ropes

-0.60

POSITIVE LOGITS

die

0.93

kus

0.92

hea

0.86

alph

0.84

enei

0.83

bard

0.81

kj

0.80

mund

0.79

fect

0.79

stein

0.79

Activations Density 0.062%

words related to political figures and news outlets

mentions of specific individuals associated with political commentary and media

No Comments

No Known Activations