INDEX

Explanations

instances of the word "who" followed by a verb or possessive pronoun

oai_token-act-pair · gpt-3.5-turbo

phrases that begin with "Anyone who" or variations that indicate a shared experience or sentiment

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

ahime

-0.81

uers

-0.70

 Tara

-0.67

chairs

-0.66

ilies

-0.65

 Neurolog

-0.64

 Braun

-0.64

romy

-0.62

ilib

-0.62

 CrossRef

-0.62

POSITIVE LOGITS

 whatsoever

1.29

 imaginable

1.04

 reasonably

0.87

 remotely

0.87

 slightest

0.86

 conceivable

0.81

astical

0.78

soever

0.76

 ever

0.76

respect

0.74

Activations Density 0.305%

instances of the word "who" followed by a verb or possessive pronoun

phrases that begin with "Anyone who" or variations that indicate a shared experience or sentiment

No Comments

No Known Activations