INDEX

Explanations

phrases related to legal or law enforcement contexts, with a particular focus on individuals

oai_token-act-pair · gpt-3.5-turbo

mentions of "person" in various contexts related to individual actions or situations

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

unctions

-0.74

 Lans

-0.71

enthal

-0.71

DL

-0.64

CCC

-0.63

 corridors

-0.63

 Lions

-0.62

rss

-0.61

urations

-0.61

 Lank

-0.61

POSITIVE LOGITS

hood

1.34

nel

1.05

who

0.84

ification

0.82

ifies

0.80

ified

0.80

uscript

0.79

 else

0.78

acles

0.77

who

0.77

Activations Density 0.032%

phrases related to legal or law enforcement contexts, with a particular focus on individuals

mentions of "person" in various contexts related to individual actions or situations

No Comments

No Known Activations