INDEX

Explanations

phrases related to deep understanding or knowledge

oai_token-act-pair · gpt-3.5-turbo

references to the concept of understanding

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

onies

-0.81

quer

-0.72

inating

-0.70

photos

-0.67

yss

-0.65

 chopping

-0.65

strip

-0.64

cake

-0.62

cow

-0.62

join

-0.61

POSITIVE LOGITS

 Understanding

0.88

 comprehension

0.84

 understanding

0.84

FUL

0.79

OF

0.78

 Understand

0.76

 thereof

0.76

ably

0.74

ledge

0.72

ually

0.72

Activations Density 0.021%

phrases related to deep understanding or knowledge

references to the concept of understanding

No Comments

No Known Activations