INDEX

Explanations

keywords related to strategic actions or plans, especially those involving controversial or deceptive methods

oai_token-act-pair · gpt-3.5-turbo

references to strategies or methods being employed in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

leased

-0.89

mbuds

-0.75

league

-0.73

worthiness

-0.72

rake

-0.69

val

-0.69

birth

-0.68

ergy

-0.65

riel

-0.65

ports

-0.64

POSITIVE LOGITS

 tactics

1.11

 tactic

0.93

 tricks

0.92

 techniques

0.79

ologies

0.77

 strategies

0.77

 lawy

0.75

 reversal

0.75

 methods

0.75

ategy

0.74

Activations Density 0.021%

keywords related to strategic actions or plans, especially those involving controversial or deceptive methods

references to strategies or methods being employed in various contexts

No Comments

No Known Activations