INDEX

Explanations

phrases related to signing or endorsing documents/agreements

oai_token-act-pair · gpt-3.5-turbo

terms related to signatories and endorsements

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

schild

-0.66

cour

-0.65

 Lands

-0.64

illet

-0.63

Shot

-0.63

ondo

-0.62

 Shots

-0.62

 Swing

-0.62

uku

-0.62

itars

-0.61

POSITIVE LOGITS

hip

0.86

ients

0.81

 thereof

0.80

Alpha

0.75

 beware

0.71

agonist

0.69

 downstream

0.66

ettings

0.66

cript

0.65

ãĤ·ãĥ£

0.63

Activations Density 0.088%

phrases related to signing or endorsing documents/agreements

terms related to signatories and endorsements

No Comments

No Known Activations