INDEX

Explanations

terms or statements that suggest implications or underlying meanings

oai_token-act-pair · gpt-3.5-turbo

phrases that convey suggestion or inference

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Steps

-0.75

Topics

-0.72

ulic

-0.68

 Bram

-0.65

Pal

-0.62

 Ready

-0.60

PG

-0.59

 Known

-0.59

 batt

-0.58

jam

-0.58

POSITIVE LOGITS

 implied

2.85

 imply

2.81

 implies

2.68

 implying

2.35

 implication

1.90

 implicitly

1.68

 insin

1.52

 inferred

1.40

 presupp

1.31

 entails

1.30

Activations Density 0.020%

terms or statements that suggest implications or underlying meanings

phrases that convey suggestion or inference

No Comments

No Known Activations