INDEX

Explanations

messages or statements with potentially significant or impactful content

oai_token-act-pair · gpt-3.5-turbo

phrases emphasizing the concept of a "message" being communicated

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

ords

-0.74

engeance

-0.74

everal

-0.71

ced

-0.71

ancies

-0.68

erenn

-0.68

endars

-0.68

umbers

-0.68

itates

-0.67

idents

-0.66

POSITIVE LOGITS

 message

1.13

 messages

1.02

 conveyed

0.91

 Messages

0.90

message

0.88

posts

0.81

 Message

0.78

FontSize

0.77

 communicated

0.77

eering

0.76

Activations Density 0.026%

messages or statements with potentially significant or impactful content

phrases emphasizing the concept of a "message" being communicated

No Comments

No Known Activations