INDEX

Explanations

phrases indicating certainty or emphasis

oai_token-act-pair · gpt-3.5-turbo

phrases asserting certainty or questioning clarity

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

ahime

-0.90

ipel

-0.81

itton

-0.80

last

-0.79

adle

-0.79

ighth

-0.77

 exting

-0.74

inse

-0.74

inki

-0.73

church

-0.72

POSITIVE LOGITS

 whatsoever

1.21

nor

0.76

REDACTED

0.71

XM

0.70

.>>

0.70

ANCE

0.69

 plainly

0.69

 mark

0.68

Width

0.67

 SHALL

0.67

Activations Density 0.031%

phrases indicating certainty or emphasis

phrases asserting certainty or questioning clarity

No Comments

No Known Activations