INDEX

Explanations

mentions of public relations-related terms

oai_token-act-pair · gpt-3.5-turbo

the acronym "PR" followed by contexts related to public relations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

lihood

-0.82

 Genesis

-0.74

 Runes

-0.70

lla

-0.69

 Resurrection

-0.67

 Presbyter

-0.66

 Tale

-0.66

 Braves

-0.65

querque

-0.64

ãĥĥãĤ¯

-0.64

POSITIVE LOGITS

INC

1.12

IME

1.02

INT

0.99

IOR

0.98

OF

0.98

ISM

0.98

IZE

0.96

OPER

0.94

ODUCT

0.93

INTER

0.92

Activations Density 0.007%

mentions of public relations-related terms

the acronym "PR" followed by contexts related to public relations

No Comments

No Known Activations