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

CHAT

-0.93

rush

-0.72

oslav

-0.71

HUD

-0.70

UM

-0.70

REAM

-0.70

moon

-0.70

LAN

-0.70

BUG

-0.69

WARN

-0.68

POSITIVE LOGITS

ãĥ¼ãĤ¯

0.71

ãĤ®

0.66

ãĥ´ãĤ¡

0.66

ãĥĺãĥ©

0.64

0.62

é¾į

0.61

eval

0.61

ãĥł

0.60

 Nicola

0.60

 horizont

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.