INDEX

Explanations

proper nouns or names

oai_token-act-pair · gpt-3.5-turbo

terms related to branding or identifying products

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

hower

-0.86

ĸļ

-0.67

Battery

-0.65

 Procedure

-0.64

 Citation

-0.63

 inconsistency

-0.60

 interstellar

-0.60

 tailor

-0.59

Þ

-0.58

 proverb

-0.58

POSITIVE LOGITS

 except

0.91

ighty

0.73

andro

0.71

together

0.70

ves

0.70

ergic

0.66

atives

0.65

table

0.64

theless

0.64

board

0.64

Activations Density 0.130%

proper nouns or names

terms related to branding or identifying products

No Comments

No Known Activations