INDEX

Explanations

instances where there are no plans to do something

oai_token-act-pair · gpt-3.5-turbo

the word "to" indicating intentions or plans

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

soDeliveryDate

-0.91

nine

-0.61

didn

-0.60

Dash

-0.60

 Lots

-0.59

 Dimension

-0.59

 Needs

-0.59

accompanied

-0.58

 Conditions

-0.56

zynski

-0.56

POSITIVE LOGITS

 anymore

1.26

 adequately

1.15

 interfere

1.15

nor

1.12

 compete

1.07

 condone

1.04

 dissu

1.02

 solve

0.99

 offend

0.99

 impede

0.98

Activations Density 0.217%

instances where there are no plans to do something

the word "to" indicating intentions or plans

No Comments

No Known Activations