INDEX

Explanations

phrases related to actions being taken or expressed by individuals

oai_token-act-pair · gpt-3.5-turbo

verbs indicating action or influence in a narrative context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

selves

-0.74

[];

-0.66

 Created

-0.66

inea

-0.65

IDs

-0.63

houses

-0.63

foundland

-0.62

[|

-0.60

 Purch

-0.60

rats

-0.60

POSITIVE LOGITS

 rhetorical

1.01

 rhet

0.96

 applause

0.93

 clarify

0.93

 sarcast

0.93

 dismissive

0.90

 sarcastic

0.89

 apologize

0.87

 apologizing

0.87

 reiter

0.87

Activations Density 0.368%

phrases related to actions being taken or expressed by individuals

verbs indicating action or influence in a narrative context

No Comments

No Known Activations