INDEX

Explanations

phrases related to small examples or elements representing a larger concept

oai_token-act-pair · gpt-3.5-turbo

references to significant issues or underlying problems

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

chev

-0.82

iors

-0.80

unctions

-0.75

Daily

-0.73

ancies

-0.72

OWN

-0.72

LY

-0.71

ummies

-0.69

lords

-0.67

cler

-0.66

POSITIVE LOGITS

 iceberg

1.39

 scale

0.87

 spear

0.82

 scales

0.82

 rope

0.78

 proverbial

0.76

 wedge

0.74

 finger

0.74

 mustard

0.72

 fingers

0.67

Activations Density 0.096%

phrases related to small examples or elements representing a larger concept

references to significant issues or underlying problems

No Comments

No Known Activations