INDEX

Explanations

phrases indicating surprise or shock

oai_token-act-pair · gpt-3.5-turbo

expressions related to shock, surprise, and alarming 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

•layer 5 - surprise

No Comments

Negative Logits

nav

-0.81

 servicing

-0.79

 bonded

-0.78

Ã¨re

-0.77

neys

-0.77

claimer

-0.76

ney

-0.75

pei

-0.74

wright

-0.74

perate

-0.73

POSITIVE LOGITS

ly

1.00

 revelations

0.92

LY

0.88

 Revelations

0.88

Trend

0.80

 revelation

0.78

Fact

0.76

 Flavoring

0.76

 statistic

0.74

 occurrences

0.72

Activations Density 0.061%

phrases indicating surprise or shock

expressions related to shock, surprise, and alarming situations

No Comments

No Known Activations