INDEX

Explanations

web addresses or email signatures containing the word "dot"

oai_token-act-pair · gpt-3.5-turbo

references to the term "dot" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

IENCE

-0.71

CVE

-0.70

episode

-0.66

FUL

-0.65

ANS

-0.65

HAM

-0.64

Listen

-0.64

committee

-0.64

 Means

-0.62

orthy

-0.62

POSITIVE LOGITS

dot

1.36

dot

1.11

olor

0.91

biz

0.87

uate

0.85

eret

0.85

hesda

0.75

roleum

0.74

roll

0.71

oday

0.71

Activations Density 0.005%

web addresses or email signatures containing the word "dot"

references to the term "dot" in various contexts

No Comments

No Known Activations