INDEX

Explanations

phrases related to approval or authorization

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "approve" and its variations in contexts related to agreements, decisions, and endorsements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ngth

-0.81

hed

-0.75

gm

-0.70

elve

-0.66

tone

-0.66

lla

-0.66

ern

-0.66

hole

-0.65

igil

-0.64

raltar

-0.64

POSITIVE LOGITS

 unanimously

0.99

 approvals

0.96

 authorizing

0.96

 waivers

0.93

 approving

0.90

 approval

0.88

 approves

0.82

 approve

0.82

animous

0.79

ably

0.77

Activations Density 0.057%

phrases related to approval or authorization

occurrences of the word "approve" and its variations in contexts related to agreements, decisions, and endorsements

No Comments

No Known Activations