INDEX

Explanations

mentions of the word "capacity" in various contexts

oai_token-act-pair · gpt-3.5-turbo

references to overall capacity

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

aida

-0.83

scene

-0.73

algia

-0.71

bart

-0.70

utorial

-0.68

nee

-0.66

cause

-0.66

roma

-0.64

esta

-0.64

erer

-0.63

POSITIVE LOGITS

 capacities

1.29

 capacity

1.17

 utilization

1.07

 Capacity

0.96

acity

0.86

abilities

0.86

capacity

0.85

ful

0.81

 limitations

0.79

 capabilities

0.79

Activations Density 0.028%

mentions of the word "capacity" in various contexts

references to overall capacity

No Comments

No Known Activations