INDEX

Explanations

phrases related to causing, preventing, or dealing with problems and challenges

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate leading or preventing negative outcomes or conflicts

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

nect

-0.77

assad

-0.76

ledged

-0.72

orld

-0.71

orest

-0.67

hest

-0.66

fig

-0.65

atri

-0.65

 Protector

-0.65

Sel

-0.64

POSITIVE LOGITS

 confusion

1.62

 misunderstand

1.52

 misunderstanding

1.41

 inconsistency

1.33

 duplication

1.27

 unnecessary

1.26

 ambiguity

1.26

 misinterpret

1.25

 clutter

1.24

 misconceptions

1.23

Activations Density 0.379%

phrases related to causing, preventing, or dealing with problems and challenges

phrases that indicate leading or preventing negative outcomes or conflicts

No Comments

No Known Activations