INDEX

Explanations

lengthy descriptions or articles

oai_token-act-pair · gpt-3.5-turbo

references to documents or discussions characterized by extensive detail or duration

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

ramid

-0.90

imating

-0.70

onom

-0.70

ery

-0.69

rollers

-0.69

ilet

-0.69

lik

-0.67

Ability

-0.67

ading

-0.66

rol

-0.65

POSITIVE LOGITS

 lengthy

0.85

 periods

0.81

 stretches

0.81

 length

0.76

 awaited

0.75

 delays

0.75

 lengths

0.74

 overdue

0.72

 queues

0.72

itud

0.71

Activations Density 0.006%

lengthy descriptions or articles

references to documents or discussions characterized by extensive detail or duration

No Comments

No Known Activations