INDEX

Explanations

proper nouns or names, especially those containing the letters "D" and "y"

oai_token-act-pair · gpt-3.5-turbo

references to specific names associated with a medical condition

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

IZE

-0.78

ãģĤ

-0.75

ä»£

-0.74

å§«

-0.73

rawdownloadcloneembedreportprint

-0.70

oice

-0.69

sburgh

-0.68

ãĤĤ

-0.67

 Austral

-0.67

 Mara

-0.65

POSITIVE LOGITS

Dy

1.21

gradation

0.95

dy

0.92

dy

0.91

rell

0.84

sty

0.81

stop

0.80

wayne

0.79

stal

0.78

grass

0.77

Activations Density 0.006%

proper nouns or names, especially those containing the letters "D" and "y"

references to specific names associated with a medical condition

No Comments

No Known Activations