INDEX

Explanations

words related to breaking, stacking, and putting pieces together

oai_token-act-pair · gpt-3.5-turbo

recurring mentions of "pieces"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Mutual

-0.66

 Predators

-0.62

 Friend

-0.59

 Lawyers

-0.58

 pronounced

-0.58

Dub

-0.58

riel

-0.57

 Monitor

-0.57

POSITIVE LOGITS

meal

1.41

 pieces

1.38

 Pieces

1.19

ngth

1.08

 piece

1.06

pieces

1.02

uania

0.90

ilage

0.89

glass

0.88

ridges

0.84

Activations Density 0.014%

words related to breaking, stacking, and putting pieces together

recurring mentions of "pieces"

No Comments

No Known Activations