INDEX

Explanations

technical terms and names related to specific projects or entities

oai_token-act-pair · gpt-3.5-turbo

proper nouns related to scientific concepts and figures, particularly in transportation and medical 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

 Defender

-0.75

rica

-0.67

roud

-0.62

 tabletop

-0.62

makers

-0.61

 Shack

-0.61

 pirates

-0.60

oice

-0.59

 Somali

-0.59

 Prosecut

-0.59

POSITIVE LOGITS

aga

0.85

aling

0.84

usalem

0.82

©¶æ¥µ

0.76

agen

0.72

als

0.72

omed

0.72

xes

0.71

aries

0.71

 Alam

0.70

Activations Density 0.048%

technical terms and names related to specific projects or entities

proper nouns related to scientific concepts and figures, particularly in transportation and medical contexts

No Comments

No Known Activations