INDEX

Explanations

terms that indicate physical proximity or close relationships

oai_token-act-pair · gpt-3.5-turbo

phrases indicating proximity or closeness

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

NZ

-0.77

cloth

-0.72

FAQ

-0.69

wine

-0.68

ONES

-0.68

orie

-0.68

ulations

-0.67

\\\\\\\\

-0.67

;;;;;;;;;;;;

-0.64

otte

-0.64

POSITIVE LOGITS

than

1.06

 than

1.04

 proximity

0.94

 scrutiny

0.92

 Than

0.92

 thereto

0.86

 examination

0.85

 resemblance

0.81

 inspection

0.80

 "$:/

0.79

Activations Density 0.018%

terms that indicate physical proximity or close relationships

phrases indicating proximity or closeness

No Comments

No Known Activations