INDEX

Explanations

phrases related to giving advice or hints

oai_token-act-pair · gpt-3.5-turbo

references to tips or pieces of advice

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

ively

-0.75

 Nationwide

-0.68

ufact

-0.68

 Recall

-0.67

Zed

-0.67

 NAME

-0.67

Mub

-0.66

effic

-0.65

 Scope

-0.64

Ago

-0.63

POSITIVE LOGITS

jar

1.06

tip

1.04

tip

0.99

 toes

0.98

ster

0.94

haps

0.92

 iceberg

0.91

sters

0.90

sy

0.88

toe

0.82

Activations Density 0.053%

phrases related to giving advice or hints

references to tips or pieces of advice

No Comments

No Known Activations