INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ctuary

-0.90

mberg

-0.81

thouse

-0.80

ulators

-0.65

alde

-0.64

idation

-0.64

 maiden

-0.63

JV

-0.63

ancies

-0.63

 fright

-0.62

POSITIVE LOGITS

 Subject

0.66

ogle

0.64

summary

0.64

Subject

0.63

 Investor

0.62

0.61

eller

0.61

INFO

0.59

ellen

0.58

opa

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.