INDEX

Explanations

sequences related to coding or encoding

oai_token-act-pair · gpt-3.5-turbo

terms related to coding and encoding processes

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

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

Negative Logits

photo

-0.71

recent

-0.71

Stra

-0.68

=-=-=-=-

-0.67

Advertisements

-0.65

Rep

-0.65

Story

-0.65

ivan

-0.64

 Athletics

-0.63

Jet

-0.63

POSITIVE LOGITS

oding

1.56

oded

1.47

ode

1.14

odes

1.13

oder

1.00

hower

0.99

ODE

0.89

orage

0.83

ependent

0.79

etermination

0.72

Activations Density 0.008%

sequences related to coding or encoding

terms related to coding and encoding processes

No Known Activations