INDEX

Explanations

technical terms and code snippets

oai_token-act-pair · gpt-3.5-turbo

references to scientific or technical concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

stre

-0.67

 relapse

-0.66

 regrets

-0.62

 constitu

-0.59

morrow

-0.59

-0.58

ndum

-0.58

 adjourn

-0.57

adiq

-0.56

Community

-0.55

POSITIVE LOGITS

Basically

0.72

 specs

0.69

 nutshell

0.69

 Basically

0.67

 Manufacturer

0.66

 Essentially

0.65

 technically

0.65

 acronym

0.65

 interchangeable

0.64

Wikipedia

0.64

Activations Density 1.252%

technical terms and code snippets

references to scientific or technical concepts

No Comments

No Known Activations