INDEX

Explanations

references to a specific abbreviation followed by a numerical value

oai_token-act-pair · gpt-3.5-turbo

specific codes or abbreviations

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

eers

-1.03

eering

-1.02

eer

-0.88

 Franch

-0.76

Downloadha

-0.71

ricular

-0.68

Loading

-0.68

croft

-0.68

CVE

-0.67

mount

-0.65

POSITIVE LOGITS

estone

0.83

itus

0.83

ying

0.76

idy

0.76

ueless

0.75

TPS

0.75

oyd

0.74

iffs

0.74

istics

0.73

ipop

0.72

Activations Density 0.057%

references to a specific abbreviation followed by a numerical value

specific codes or abbreviations

No Comments

No Known Activations