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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

LESS

-0.92

 Caption

-0.73

BuyableInstoreAndOnline

-0.70

pat

-0.69

 condem

-0.69

 sqor

-0.68

disabled

-0.66

 Dwell

-0.66

REF

-0.65

éĸ

-0.64

POSITIVE LOGITS

ventures

0.84

estyles

0.79

igree

0.77

iltration

0.74

pmwiki

0.71

ittees

0.67

ignt

0.66

essert

0.66

issions

0.65

iphate

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.