INDEX

Explanations

phrases related to posing questions or challenges

oai_token-act-pair · gpt-3.5-turbo

instances of the word "posed" indicating a question or a challenge

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Init

-0.62

tery

-0.59

FIR

-0.59

 allot

-0.57

aware

-0.57

 memory

-0.57

zos

-0.57

 Memory

-0.56

 deaf

-0.54

 Switch

-0.54

POSITIVE LOGITS

 posed

3.76

 posing

2.13

 poses

2.09

 pose

1.99

 presented

1.27

 faced

1.26

 fielded

1.23

 risked

1.20

 provoked

1.13

 wielded

1.09

Activations Density 0.008%

phrases related to posing questions or challenges

instances of the word "posed" indicating a question or a challenge

No Comments

No Known Activations