INDEX

Explanations

words related to physical or metaphorical impairment or immobilization

oai_token-act-pair · gpt-3.5-turbo

terminology related to physical incapacitation or dysfunction

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Ryder

-0.70

rons

-0.70

atican

-0.69

rer

-0.69

endum

-0.68

lli

-0.67

rite

-0.67

LR

-0.66

Values

-0.66

videos

-0.66

POSITIVE LOGITS

 paralyzed

1.03

 paraly

1.03

 crippled

0.99

locked

0.95

 cripp

0.93

 halting

0.82

 paralysis

0.80

luaj

0.77

 limp

0.77

 blinded

0.76

Activations Density 0.031%

words related to physical or metaphorical impairment or immobilization

terminology related to physical incapacitation or dysfunction

No Comments

No Known Activations