INDEX

Explanations

phrases or sentence endings enclosed in parentheses

oai_token-act-pair · gpt-3.5-turbo

puncuation marks, specifically closed parentheses and their surrounding contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 cour

-0.63

tip

-0.62

lish

-0.61

 flakes

-0.57

 proper

-0.56

 waste

-0.55

 binge

-0.55

 tenant

-0.55

 grit

-0.54

tons

-0.54

POSITIVE LOGITS

pmwiki

0.97

 âĵĺ

0.90

posted

0.87

 Unloaded

0.86

=>

0.82

[+

0.78

ategories

0.78

BILITIES

0.76

 âĶľâĶĢâĶĢ

0.75

 Anyway

0.75

Activations Density 0.078%

phrases or sentence endings enclosed in parentheses

puncuation marks, specifically closed parentheses and their surrounding contexts

No Comments

No Known Activations