INDEX

Explanations

names and titles starting with the letter "A" or "R"

oai_token-act-pair · gpt-3.5-turbo

specific names or terms, particularly proper nouns or notable individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

Reviewer

-0.78

Track

-0.73

 unnecess

-0.61

Inc

-0.59

Parameter

-0.59

tail

-0.58

killer

-0.57

 mockery

-0.57

REC

-0.56

mington

-0.56

POSITIVE LOGITS

 Pradesh

1.04

heim

0.89

neau

0.79

henko

0.75

ndra

0.72

uez

0.69

ault

0.68

ergic

0.67

ð

0.66

abeth

0.65

Activations Density 0.060%

names and titles starting with the letter "A" or "R"

specific names or terms, particularly proper nouns or notable individuals

No Comments

No Known Activations