INDEX

Explanations

locations or places

oai_token-act-pair · gpt-3.5-turbo

references to specific locations or places

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ERAL

-0.68

uded

-0.67

natureconservancy

-0.66

agascar

-0.66

iley

-0.63

unci

-0.63

CHAT

-0.63

uding

-0.63

olls

-0.61

tnc

-0.61

POSITIVE LOGITS

holder

1.21

holders

1.11

bos

0.96

ername

0.91

 Place

0.88

forth

0.87

bo

0.83

Place

0.82

down

0.78

place

0.75

Activations Density 0.034%

locations or places

references to specific locations or places

No Comments

No Known Activations