INDEX

Explanations

phrases related to public relations

oai_token-act-pair · gpt-3.5-turbo

references to public relations concepts and entities

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

\\\\\\\\\\\\\\\\

-0.81

ÅŁ

-0.78

ropolis

-0.74

YC

-0.73

 Sicily

-0.73

asar

-0.73

âĶĢâĶĢâĶĢâĶĢâĶĢâĶĢâĶĢâĶĢ

-0.71

erton

-0.69

prus

-0.68

 Schne

-0.68

POSITIVE LOGITS

hips

1.57

hip

0.98

pring

0.90

 relations

0.86

 warr

0.76

etter

0.74

bending

0.74

hops

0.73

relations

0.70

 diagrams

0.69

Activations Density 0.020%

phrases related to public relations

references to public relations concepts and entities

No Comments

No Known Activations