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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

MED

-0.79

key

-0.75

ergic

-0.74

ICAL

-0.72

written

-0.72

nia

-0.70

 nurs

-0.69

versely

-0.65

itor

-0.65

isance

-0.64

POSITIVE LOGITS

0.84

pic

0.81

 congr

0.77

 applause

0.73

.#

0.70

 blacks

0.69

outer

0.67

 Olivier

0.67

 rainbow

0.65

(#

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.