INDEX

Explanations

conjunctions connecting different elements or concepts

oai_token-act-pair · gpt-3.5-turbo

phrases emphasizing contrast or exceptions

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

coat

-0.76

orc

-0.75

irm

-0.73

ct

-0.71

wake

-0.71

velop

-0.70

oresc

-0.67

wrap

-0.67

ctor

-0.66

Plex

-0.66

POSITIVE LOGITS

 also

1.00

 ALSO

0.86

 actively

0.81

 secondly

0.79

 importantly

0.75

 downright

0.75

 moreover

0.74

yon

0.71

 possibly

0.71

also

0.70

Activations Density 0.051%

conjunctions connecting different elements or concepts

phrases emphasizing contrast or exceptions

No Comments

No Known Activations