INDEX

Explanations

words related to the core concept or essential components of various topics

oai_token-act-pair · gpt-3.5-turbo

keywords related to core principles, values, and mechanics in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

quished

-0.72

 compr

-0.72

 unwitting

-0.67

enne

-0.67

uterte

-0.66

EStream

-0.65

inea

-0.65

nesday

-0.65

Ń·

-0.64

seys

-0.62

POSITIVE LOGITS

 fundamentals

0.86

 tenets

0.84

 literacy

0.78

 curriculum

0.75

gebra

0.72

 foundational

0.70

 core

0.69

 fundament

0.69

 stabilization

0.69

 Knowledge

0.68

Activations Density 0.215%

words related to the core concept or essential components of various topics

keywords related to core principles, values, and mechanics in various contexts

No Comments

No Known Activations