INDEX

Explanations

images of predictability or patterns

oai_token-act-pair · gpt-3.5-turbo

references to consistency or predictability

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

gur

-0.95

sten

-0.82

inis

-0.80

onia

-0.78

tein

-0.77

athy

-0.76

atha

-0.75

zanne

-0.75

gar

-0.73

inth

-0.72

POSITIVE LOGITS

 predictable

0.97

ãĥĻ

0.80

 unpredict

0.78

 adolesc

0.78

é¾

0.75

 juven

0.75

 surpr

0.74

 Predict

0.74

 disson

0.73

aneously

0.73

Activations Density 0.006%

images of predictability or patterns

references to consistency or predictability

No Comments

No Known Activations