INDEX

Explanations

sentences discussing statements or quotes, often including opinions or analysis

oai_token-act-pair · gpt-3.5-turbo

quotes or dialogue in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 scrimmage

-0.82

 honors

-0.82

 honored

-0.81

 honor

-0.80

 gracious

-0.76

 honorable

-0.75

 ceremon

-0.74

 swore

-0.71

TMZ

-0.71

 goof

-0.71

POSITIVE LOGITS

Incre

1.20

Therefore

1.07

 Increasing

1.07

Moreover

1.05

Furthermore

1.03

Understanding

1.02

Currently

1.00

Indeed

0.98

Ultimately

0.98

Companies

0.95

Activations Density 0.715%

sentences discussing statements or quotes, often including opinions or analysis

quotes or dialogue in the text

No Comments

No Known Activations