INDEX

Explanations

references to different abilities or capabilities

oai_token-act-pair · gpt-3.5-turbo

phrases related to capabilities or competencies

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 Surv

-0.67

gar

-0.66

adish

-0.63

anski

-0.63

roman

-0.62

 Observ

-0.62

 Wrap

-0.61

gone

-0.61

 Bride

-0.61

 Coconut

-0.61

POSITIVE LOGITS

 destro

0.90

auga

0.87

ibilities

0.86

ibility

0.86

Ability

0.85

 impaired

0.83

 ability

0.82

Reviewer

0.77

 incap

0.75

bodied

0.74

Activations Density 0.030%

references to different abilities or capabilities

phrases related to capabilities or competencies

No Comments

No Known Activations