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

stal

-0.82

spect

-0.77

 calib

-0.73

olars

-0.69

clud

-0.68

itaire

-0.67

 domestically

-0.66

oppable

-0.66

itiz

-0.66

schild

-0.64

POSITIVE LOGITS

atha

0.64

 Ballard

0.63

row

0.62

 abiding

0.61

ATH

0.60

RESULTS

0.59

 endif

0.59

 bond

0.59

 directors

0.58

APD

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.