INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

natureconservancy

-0.77

Sharp

-0.69

aspx

-0.66

oren

-0.65

 habitat

-0.61

methyl

-0.59

 poles

-0.59

hyp

-0.59

 pole

-0.58

 eleph

-0.57

POSITIVE LOGITS

ĸļ

0.72

LET

0.70

Frames

0.65

asion

0.63

Trend

0.63

 Liter

0.62

say

0.61

 psychiat

0.61

rite

0.59

landers

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.