INDEX

Explanations

time-related expressions and indications

oai_token-act-pair · gpt-3.5-turbo

phrases indicating familiarity or prior knowledge about a topic

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Dialogue

-0.63

 prolong

-0.63

 sacrific

-0.58

 Hispan

-0.58

 Spectre

-0.58

grades

-0.57

INGS

-0.56

 downgrade

-0.56

 Akin

-0.55

 inflic

-0.54

POSITIVE LOGITS

 know

1.15

 noticed

1.13

know

1.11

 guessed

1.10

 probably

1.09

 heard

1.03

 familiar

1.01

 already

0.98

 undoubtedly

0.96

 knew

0.95

Activations Density 0.193%

time-related expressions and indications

phrases indicating familiarity or prior knowledge about a topic

No Comments

No Known Activations