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

ttes

-0.73

 Behavioral

-0.69

artments

-0.68

enaries

-0.68

acters

-0.67

 Decker

-0.67

glers

-0.66

tails

-0.65

chens

-0.64

tsky

-0.64

POSITIVE LOGITS

0.89

 inclination

0.75

Pg

0.71

APH

0.69

 pinch

0.68

 oblig

0.67

err

0.66

 latitude

0.66

 resil

0.65

phrine

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.