INDEX

Explanations

phrases related to being off course or deviating from expectations

oai_token-act-pair · gpt-3.5-turbo

phrases related to media references or programming content

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

CVE

-0.76

RIC

-0.71

ulhu

-0.70

afia

-0.69

non

-0.66

å°Ĩ

-0.65

VEN

-0.65

CLA

-0.64

ruary

-0.63

shapeshifter

-0.62

POSITIVE LOGITS

 beaten

0.98

 leash

0.97

 coasts

0.94

 hinges

0.92

 coast

0.89

 island

0.89

 rails

0.88

 heels

0.87

 shelves

0.86

 shores

0.85

Activations Density 0.045%

phrases related to being off course or deviating from expectations

phrases related to media references or programming content

No Comments

No Known Activations