INDEX

Explanations

specific encoded strings within the text

oai_token-act-pair · gpt-3.5-turbo

the letter "Ļ" in varying contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 imitation

-0.73

 Seym

-0.66

 disadvant

-0.63

 Conrad

-0.62

 Presbyter

-0.62

 Raleigh

-0.61

 Gutenberg

-0.59

 carbohyd

-0.59

 Wichita

-0.59

 deepening

-0.59

POSITIVE LOGITS

ï¸ı

1.10

Â·Â·

0.86

å¥³

0.84

âĶĢâĶĢ

0.82

mt

0.82

Â£

0.81

ï¸

0.78

Lens

0.76

ution

0.75

ATH

0.75

Activations Density 0.600%

specific encoded strings within the text

the letter "Ļ" in varying contexts

No Comments

No Known Activations