INDEX

Explanations

phrases related to lessons or advice

oai_token-act-pair · gpt-3.5-turbo

iterations of the word "less" in a text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

AE

-0.72

mberg

-0.67

 Dresden

-0.63

 Unlimited

-0.62

 sqor

-0.61

cano

-0.59

å¯

-0.58

 Lobby

-0.57

create

-0.57

[|

-0.55

POSITIVE LOGITS

 than

1.22

ened

1.16

ening

1.08

 fortunate

0.97

ons

0.95

ens

0.94

than

0.94

 traveled

0.90

 Than

0.90

 travelled

0.84

Activations Density 0.043%

phrases related to lessons or advice

iterations of the word "less" in a text

No Comments

No Known Activations