INDEX

Explanations

adverbs or words indicating a transformation or explanation

oai_token-act-pair · gpt-3.5-turbo

phrases that convey the concept of being fundamental or critical in nature

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

-0.84

particularly

-0.80

eful

-0.73

 particularly

-0.71

Others

-0.69

Leary

-0.69

OF

-0.68

Ey

-0.68

efully

-0.68

acious

-0.67

POSITIVE LOGITS

 boils

0.86

 worthless

0.81

 starve

0.81

 unlimited

0.79

 meaningless

0.79

 indistinguishable

0.79

 unchanged

0.78

 doubling

0.78

 useless

0.76

 equivalent

0.76

Activations Density 0.067%

adverbs or words indicating a transformation or explanation

phrases that convey the concept of being fundamental or critical in nature

No Comments

No Known Activations