INDEX

Explanations

references to the act of reducing or decreasing something, particularly in the context of rates or standards

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

instances of the word "lowered"

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

Asia

-0.75

icity

-0.74

POL

-0.72

Phill

-0.72

ious

-0.71

present

-0.71

SPONSORED

-0.71

 Samoa

-0.70

 Origins

-0.67

Pros

-0.66

POSITIVE LOGITS

 lowering

1.09

 inhib

0.94

baugh

0.94

imar

0.92

 lowered

0.89

 benches

0.87

aution

0.85

 carbohyd

0.83

hetically

0.81

 priced

0.78

Activations Density 0.008%

references to the act of reducing or decreasing something, particularly in the context of rates or standards

instances of the word "lowered"

No Comments

No Known Activations