INDEX

Explanations

phrases indicating that something is significantly far from a location or outcome

oai_token-act-pair · gpt-3.5-turbo

phrases indicating a lack of options or direction

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

 Reef

-0.78

ilt

-0.68

 mash

-0.67

tein

-0.64

atto

-0.60

uno

-0.60

eds

-0.60

osterone

-0.59

apo

-0.59

ivation

-0.59

POSITIVE LOGITS

 else

1.36

 Else

1.02

Else

0.97

abouts

0.93

near

0.87

 Near

0.87

 near

0.85

else

0.80

 nearer

0.75

upon

0.74

Activations Density 0.025%

phrases indicating that something is significantly far from a location or outcome

phrases indicating a lack of options or direction

No Comments

No Known Activations