INDEX

Explanations

words related to truth and deception

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

instances of the concept of truth and lies

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Instrument

-0.67

ERC

-0.66

 Pharm

-0.66

 Instr

-0.66

Sto

-0.65

 Spice

-0.62

 Rivals

-0.61

 Turk

-0.61

oned

-0.61

 elig

-0.60

POSITIVE LOGITS

fulness

1.45

fully

1.09

telling

1.05

 serum

1.02

iness

1.00

about

0.97

ulence

0.92

 about

0.91

seekers

0.91

behind

0.85

Activations Density 0.059%

words related to truth and deception

instances of the concept of truth and lies

No Comments

No Known Activations