INDEX

Explanations

words related to dishonesty and unethical behavior

oai_token-act-pair · gpt-3.5-turbo

terms related to deceitful or unethical behavior

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

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

/+

-0.78

anza

-0.78

izu

-0.75

iph

-0.72

xit

-0.71

ENDED

-0.71

pler

-0.69

iphate

-0.69

VILLE

-0.69

vu

-0.69

POSITIVE LOGITS

 shady

1.28

 dealings

1.11

 crooked

0.91

 unethical

0.85

 undermin

0.82

 broker

0.80

kefeller

0.80

abouts

0.77

 corners

0.77

 subcontract

0.76

Activations Density 0.014%

words related to dishonesty and unethical behavior

terms related to deceitful or unethical behavior

No Comments

No Known Activations