INDEX

Explanations

words related to rankings or listings

oai_token-act-pair · gpt-3.5-turbo

conjunctions and transitional phrases

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

xual

-0.71

erent

-0.69

iot

-0.67

rolet

-0.65

oros

-0.65

opathic

-0.65

ellig

-0.64

ribing

-0.64

²¾

-0.63

ctions

-0.61

POSITIVE LOGITS

 arguably

1.00

 certainly

0.90

 fourth

0.88

 fifth

0.87

 sixth

0.87

 ninth

0.86

 secondly

0.82

 unmatched

0.81

 longest

0.81

 surpass

0.81

Activations Density 0.225%

words related to rankings or listings

conjunctions and transitional phrases

No Comments

No Known Activations