INDEX

Explanations

words or suffixes related to descriptions or characteristics

oai_token-act-pair · gpt-3.5-turbo

technical terms related to descriptions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Howell

-0.79

 Charlie

-0.69

 Hawking

-0.68

moon

-0.65

 Nebula

-0.65

 grandchildren

-0.65

Moon

-0.64

 Carey

-0.63

claw

-0.62

 Hole

-0.62

POSITIVE LOGITS

ript

5.08

raph

1.31

bps

1.26

CRIP

1.16

rypt

1.11

ription

1.10

ibi

0.92

ipt

0.91

ruct

0.86

ribing

0.85

Activations Density 0.031%

words or suffixes related to descriptions or characteristics

technical terms related to descriptions

No Comments

No Known Activations