INDEX

Explanations

phrases related to movies or cultural productions

oai_token-act-pair · gpt-3.5-turbo

phrases and punctuation related to titles or names

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

worldly

-0.61

 goodwill

-0.59

ãĥ¬

-0.57

 eternity

-0.57

stroke

-0.54

 sergeant

-0.53

itant

-0.53

 Salary

-0.51

iquette

-0.51

hp

-0.50

POSITIVE LOGITS

 which

1.37

which

1.16

 wherein

1.08

 whereby

0.90

 published

0.88

 featuring

0.88

 billed

0.84

 whose

0.84

 where

0.84

 released

0.83

Activations Density 0.234%

phrases related to movies or cultural productions

phrases and punctuation related to titles or names

No Comments

No Known Activations