INDEX

Explanations

words related to things that are widely known or used

oai_token-act-pair · gpt-3.5-turbo

instances of the word "popular."

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

thur

-0.76

agher

-0.74

ignt

-0.73

heed

-0.73

oti

-0.70

 clad

-0.69

ermott

-0.67

aca

-0.67

ural

-0.66

fter

-0.65

POSITIVE LOGITS

ized

0.82

popular

0.82

ised

0.80

izing

0.79

izations

0.77

isations

0.76

 enterprises

0.73

izer

0.72

 favorites

0.70

ity

0.70

Activations Density 0.021%

words related to things that are widely known or used

instances of the word "popular."

No Comments

No Known Activations