INDEX

Explanations

phrases related to information dissemination or communication

oai_token-act-pair · gpt-3.5-turbo

references to consumer-related topics and trends

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

Tour

-0.55

ufact

-0.54

Kid

-0.54

Sold

-0.54

Shop

-0.52

Pic

-0.51

Gallery

-0.50

Movie

-0.50

VIDEO

-0.50

Inside

-0.49

POSITIVE LOGITS

 presupp

0.88

 fallacy

0.80

 methodological

0.78

 subjective

0.76

 causation

0.75

 arise

0.74

 implicit

0.74

 inherent

0.71

 heterogeneity

0.71

 arises

0.70

Activations Density 0.724%

phrases related to information dissemination or communication

references to consumer-related topics and trends

No Comments

No Known Activations