INDEX

Explanations

words related to specific locations or geographic regions

oai_token-act-pair · gpt-3.5-turbo

references to various islands and island groups

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

apers

-0.86

aper

-0.76

aped

-0.75

jee

-0.75

 Hoover

-0.69

kus

-0.68

rano

-0.68

nce

-0.66

nea

-0.66

acher

-0.65

POSITIVE LOGITS

 Islands

1.52

 islands

0.95

 Island

0.80

 Strait

0.77

hell

0.75

 Graveyard

0.74

 Albania

0.72

ylum

0.72

rift

0.71

 island

0.70

Activations Density 0.011%

words related to specific locations or geographic regions

references to various islands and island groups

No Comments

No Known Activations