INDEX

Explanations

items labeled with a specific numbering format, likely in a list

oai_token-act-pair · gpt-3.5-turbo

numerical values and specifications related to technology and products

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

netflix

-0.73

rely

-0.72

TextColor

-0.69

deen

-0.63

stood

-0.62

 Uriel

-0.60

ufact

-0.60

 tradem

-0.59

inic

-0.59

rette

-0.58

POSITIVE LOGITS

acity

1.12

 Downing

0.94

^{

0.91

th

0.90

tons

0.88

%"

0.87

0.86

0.84

Activations Density 0.075%

items labeled with a specific numbering format, likely in a list

numerical values and specifications related to technology and products

No Comments

No Known Activations