INDEX

Explanations

ones with non-zero activation values

oai_token-act-pair · gpt-3.5-turbo

elements related to coding or programming concepts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 eleph

-0.92

hement

-0.73

ÃĥÃĤÃĥÃĤ

-0.70

 newcom

-0.69

 pione

-0.67

aditional

-0.66

 occas

-0.66

 proport

-0.64

 Burnett

-0.64

 undermin

-0.62

POSITIVE LOGITS

Requirements

0.98

Testing

0.93

↵

0.92

Async

0.91

ÂłÂłÂł

0.91

github

0.90

Installation

0.88

Deploy

0.84

package

0.84

Usage

0.84

Activations Density 0.266%

ones with non-zero activation values

elements related to coding or programming concepts

No Comments

No Known Activations