INDEX

Explanations

references to specific political groups or individuals, specifically related to the acronym "PTI."

oai_token-act-pair · gpt-3.5-turbo

references to "PT" and its various contexts

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

hold

-0.82

flies

-0.81

mble

-0.80

perse

-0.79

ufact

-0.74

bars

-0.73

cle

-0.71

cles

-0.70

bye

-0.69

xual

-0.66

POSITIVE LOGITS

 Cruiser

0.94

WS

0.87

ypes

0.84

TP

0.81

IME

0.79

ION

0.79

BN

0.78

EMBER

0.77

het

0.77

raction

0.76

Activations Density 0.012%

references to specific political groups or individuals, specifically related to the acronym "PTI."

references to "PT" and its various contexts

No Comments

No Known Activations