INDEX

Explanations

action words related to avoidance and escape from consequences or responsibilities

oai_token-act-pair · gpt-3.5-turbo

phrases related to evasion or avoidance of consequences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 wisely

-0.69

Altern

-0.65

Favorite

-0.64

Textures

-0.64

intosh

-0.62

pring

-0.60

apsed

-0.58

gif

-0.58

 Finished

-0.58

ificent

-0.57

POSITIVE LOGITS

 scrutiny

1.34

 penalties

1.25

 restrictions

1.20

 wrath

1.18

 pitfalls

1.18

 cens

1.12

 prosecution

1.12

 prohibitions

1.09

 penalty

1.08

 repercussions

1.04

Activations Density 0.312%

action words related to avoidance and escape from consequences or responsibilities

phrases related to evasion or avoidance of consequences

No Comments

No Known Activations