INDEX

Explanations

sequences or lists of items or events

oai_token-act-pair · gpt-3.5-turbo

phrases indicating sequences or ordered lists

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

tek

-0.92

WT

-0.76

WI

-0.72

BAT

-0.68

Fax

-0.67

 Valkyrie

-0.67

tics

-0.65

ibur

-0.65

tech

-0.65

mit

-0.65

POSITIVE LOGITS

 chronological

1.35

 alphabet

1.28

 comma

1.14

 descending

1.06

 consecut

1.05

 ascending

1.03

 overlapping

0.98

phabet

0.97

 geographically

0.96

 numbered

0.95

Activations Density 0.327%

sequences or lists of items or events

phrases indicating sequences or ordered lists

No Comments

No Known Activations