INDEX

Explanations

verbs indicating action in a negative context

oai_token-act-pair · gpt-3.5-turbo

verbs indicating absence, failure, or lack of action

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

ortment

-0.91

soDeliveryDate

-0.76

 largeDownload

-0.71

tnc

-0.68

assetsadobe

-0.67

itton

-0.66

unknown

-0.65

figured

-0.65

soType

-0.64

nown

-0.64

POSITIVE LOGITS

 anymore

1.49

any

1.10

 anything

1.06

nor

1.06

 anybody

1.04

 anywhere

1.02

 necessarily

0.99

 slightest

0.97

 whatsoever

0.89

 anyone

0.89

Activations Density 0.249%

verbs indicating action in a negative context

verbs indicating absence, failure, or lack of action

No Comments

No Known Activations