INDEX

Explanations

phrases related to providing information or recounting events

oai_token-act-pair · gpt-3.5-turbo

phrases related to conflicting statements or contradictory information

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

 pioneered

-0.86

 abolished

-0.77

 Designed

-0.74

 pioneers

-0.73

 thriving

-0.72

 redesigned

-0.71

 Pione

-0.69

 rebuilt

-0.69

requires

-0.69

 Beaut

-0.68

POSITIVE LOGITS

 conversation

1.41

 reply

1.30

 misunderstanding

1.26

 remark

1.22

 comment

1.21

 accusation

1.21

 verbal

1.20

 questioning

1.19

 conversations

1.19

 explanation

1.17

Activations Density 0.794%

phrases related to providing information or recounting events

phrases related to conflicting statements or contradictory information

No Comments

No Known Activations