INDEX

Explanations

keywords related to a process or event that usually follows an initial action

oai_token-act-pair · gpt-3.5-turbo

references to repetitive actions or occurrences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

REE

-0.78

rug

-0.76

 âĸº

-0.70

ander

-0.69

Ø³

-0.69

æ©

-0.68

Ø§ÙĦ

-0.68

âĶĢâĶĢâĶĢâĶĢâĶĢâĶĢâĶĢâĶĢ

-0.67

NRS

-0.66

NR

-0.65

POSITIVE LOGITS

 Upon

0.83

 satisfied

0.77

ward

0.73

 bitten

0.70

 disembark

0.68

 sworn

0.67

 again

0.65

 booted

0.64

 recovering

0.63

 upon

0.63

Activations Density 0.021%

keywords related to a process or event that usually follows an initial action

references to repetitive actions or occurrences

No Comments

No Known Activations