INDEX

Explanations

the word "ci" with high activations

oai_token-act-pair · gpt-3.5-turbo

references to the field of science and its sub-disciplines

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

locked

-0.74

stage

-0.69

stood

-0.68

rooms

-0.67

ãĥī

-0.65

 Malays

-0.64

lain

-0.64

 GOODMAN

-0.64

boards

-0.63

ton

-0.62

POSITIVE LOGITS

Ã¨re

0.95

ences

0.92

enza

0.91

zona

0.86

ennes

0.86

encia

0.84

aches

0.83

otti

0.83

pe

0.83

obile

0.82

Activations Density 0.022%

the word "ci" with high activations

references to the field of science and its sub-disciplines

No Comments

No Known Activations