INDEX

Explanations

words related to medical conditions or treatments

oai_token-act-pair · gpt-3.5-turbo

references to magic or magical themes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 WARN

-0.68

elled

-0.63

essee

-0.61

SOURCE

-0.59

chie

-0.58

earned

-0.57

 taller

-0.57

 sniff

-0.57

Bre

-0.56

 senior

-0.56

POSITIVE LOGITS

agic

1.12

rament

1.03

otine

0.90

ãĥĥãĤ¯

0.86

Pwr

0.86

istics

0.83

 è£ıè

0.80

cum

0.79

azine

0.79

mson

0.76

Activations Density 0.004%

words related to medical conditions or treatments

references to magic or magical themes

No Comments

No Known Activations