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

 Garn

-0.73

 ingred

-0.73

 prospect

-0.65

 retrieval

-0.64

 consultations

-0.63

USE

-0.63

 planner

-0.61

 vocabulary

-0.61

 grocery

-0.61

nep

-0.58

POSITIVE LOGITS

inelli

0.78

 msec

0.75

ricanes

0.69

elf

0.69

usky

0.69

mask

0.68

swick

0.67

kamp

0.67

paio

0.66

zag

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.