INDEX

Explanations

words related to advertisements like "poster" and "bragging"

oai_token-act-pair · gpt-3.5-turbo

references to posters and related imagery

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

estial

-0.81

Ago

-0.78

%]

-0.77

ESSION

-0.74

hews

-0.73

owship

-0.71

IELD

-0.67

 Liberties

-0.67

IVE

-0.67

efe

-0.66

POSITIVE LOGITS

 poster

1.07

 posters

0.99

iors

0.88

 flyer

0.84

onymous

0.81

 Poster

0.79

ieu

0.79

pillar

0.77

ity

0.73

 flyers

0.72

Activations Density 0.014%

words related to advertisements like "poster" and "bragging"

references to posters and related imagery

No Comments

No Known Activations