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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

edin

-0.75

thia

-0.66

eff

-0.64

 Authority

-0.64

 Allaah

-0.63

 Cabinet

-0.61

 Immunity

-0.61

 Rapp

-0.61

 Honolulu

-0.61

="#

-0.61

POSITIVE LOGITS

ovie

0.75

artz

0.71

obook

0.67

analysis

0.66

agonist

0.66

oppers

0.62

erv

0.62

 Quentin

0.62

Ã©n

0.62

yll

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.