INDEX

Explanations

newly introduced terms or concepts

oai_token-act-pair · gpt-3.5-turbo

mentions of the word "new."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

ality

-0.67

bey

-0.64

nikov

-0.63

zzi

-0.61

-0.60

bys

-0.60

endez

-0.59

rior

-0.59

antry

-0.59

hunt

-0.58

POSITIVE LOGITS

new

3.15

new

2.32

 newly

1.99

 newest

1.86

 newer

1.74

 newfound

1.71

NEW

1.70

 newcomers

1.55

NEW

1.48

 fresh

1.46

Activations Density 0.102%

newly introduced terms or concepts

mentions of the word "new."

No Comments

No Known Activations