INDEX

Explanations

stamp-related terms or actions

oai_token-act-pair · gpt-3.5-turbo

references to postage stamps

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

MpServer

-0.78

 traged

-0.77

 Stras

-0.75

ctory

-0.73

 Aval

-0.72

izons

-0.71

Kut

-0.68

thodox

-0.66

scill

-0.66

pheus

-0.65

POSITIVE LOGITS

eding

0.98

 stamps

0.92

 stamp

0.85

itious

0.73

cards

0.73

ochet

0.71

erer

0.70

utation

0.69

uity

0.69

oring

0.69

Activations Density 0.016%

stamp-related terms or actions

references to postage stamps

No Comments

No Known Activations