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

sembly

-0.76

 Cand

-0.64

 Rican

-0.62

 Brief

-0.61

idates

-0.61

Tik

-0.61

atel

-0.61

OPE

-0.59

 Elections

-0.59

area

-0.57

POSITIVE LOGITS

çīĪ

0.79

RED

0.79

alg

0.76

GU

0.75

Reloaded

0.74

ellation

0.70

RW

0.69

rg

0.69

LU

0.68

INGTON

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.