INDEX

Explanations

words related to consistency and being consistent with rules or values

oai_token-act-pair · gpt-3.5-turbo

references to consistency in policies, practices, or behaviors

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

doors

-0.82

olit

-0.79

worms

-0.75

mong

-0.75

crow

-0.73

doms

-0.72

thur

-0.71

stals

-0.70

tu

-0.70

rection

-0.69

POSITIVE LOGITS

 consistency

0.88

itarian

0.88

ibilities

0.83

 offender

0.82

iated

0.81

ually

0.79

ibly

0.78

icut

0.78

aneously

0.77

ively

0.77

Activations Density 0.033%

words related to consistency and being consistent with rules or values

references to consistency in policies, practices, or behaviors

No Comments

No Known Activations