INDEX

Explanations

strings that resemble technical identifiers or codes

oai_token-act-pair · gpt-3.5-turbo

structured data and identifiers, particularly related to numbers and classifications

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

ritch

-0.84

bush

-0.83

atan

-0.83

issance

-0.83

sonian

-0.79

pless

-0.79

imar

-0.79

uce

-0.76

ersive

-0.75

achev

-0.74

POSITIVE LOGITS

 redacted

0.77

 Description

0.76

 annotation

0.71

Mine

0.71

coded

0.70

 informant

0.69

 numbering

0.68

 registration

0.67

 NAME

0.66

tag

0.66

Activations Density 0.275%

strings that resemble technical identifiers or codes

structured data and identifiers, particularly related to numbers and classifications

No Comments

No Known Activations