INDEX

Explanations

mathematical notation symbols and code snippets

oai_token-act-pair · gpt-3.5-turbo

patterns resembling code syntax or mathematical notation

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

 Chambers

-0.75

 poppy

-0.72

Scar

-0.67

 Balk

-0.67

patch

-0.66

 Garrison

-0.62

 Americas

-0.61

Kay

-0.61

 Bomber

-0.61

hops

-0.60

POSITIVE LOGITS

_{

3.70

^{

3.24

{\

2.55

{{

2.40

{\

2.36

{{

2.04

${

1.89

({

1.89

={

1.79

${

1.74

Activations Density 0.028%

mathematical notation symbols and code snippets

patterns resembling code syntax or mathematical notation

No Comments

No Known Activations