INDEX

Explanations

terms related to high temperature or intense activity

oai_token-act-pair · gpt-3.5-turbo

references to the word "hot" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 guiActiveUn

-0.80

ufact

-0.80

ĸļ

-0.78

uther

-0.75

yss

-0.72

ajor

-0.72

 conclud

-0.70

uth

-0.70

©¶æ

-0.67

ngth

-0.66

POSITIVE LOGITS

dogs

0.99

 spots

0.96

bed

0.95

headed

0.94

dog

0.93

ness

0.91

shots

0.89

water

0.88

shot

0.87

ened

0.86

Activations Density 0.025%

terms related to high temperature or intense activity

references to the word "hot" in various contexts

No Comments

No Known Activations