INDEX

Explanations

phrases related to intensity or degree, often indicating strength or conclusiveness

oai_token-act-pair · gpt-3.5-turbo

references to the number ten in various contexts

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

 DRAG

-0.68

 jerk

-0.62

TextColor

-0.61

****************

-0.60

HY

-0.60

à¼

-0.59

LCS

-0.59

 Indust

-0.58

Reviewer

-0.57

Activity

-0.56

POSITIVE LOGITS

acity

1.61

aciously

1.54

acious

1.51

uous

1.44

uously

1.34

ured

1.33

ements

1.19

ancies

1.14

orio

1.10

 thousand

1.06

Activations Density 0.035%

phrases related to intensity or degree, often indicating strength or conclusiveness

references to the number ten in various contexts

No Comments

No Known Activations