INDEX

Explanations

No Explanations Found

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

Quote

-0.74

________________________________

-0.72

 Mysteries

-0.72

 Error

-0.72

 Lies

-0.72

steps

-0.72

redients

-0.72

 Problems

-0.70

Error

-0.70

eals

-0.69

POSITIVE LOGITS

agher

0.81

 pitch

0.69

©¶æ

0.66

 compose

0.65

 monop

0.65

 dock

0.63

 outfit

0.63

 helm

0.63

 scatter

0.63

 blanket

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.