INDEX

Explanations

legal or regulatory terms or contexts

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "subject" and its variations in different contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ERROR

-0.70

FLAG

-0.66

 Zimmer

-0.64

 straw

-0.64

BLIC

-0.62

LOT

-0.62

IELD

-0.61

CT

-0.61

EEK

-0.61

 Alps

-0.61

POSITIVE LOGITS

ively

0.95

ivity

0.94

matter

0.91

itatively

0.90

 Subjects

0.88

ivist

0.88

subject

0.82

ivities

0.81

icals

0.81

ically

0.80

Activations Density 0.012%

legal or regulatory terms or contexts

occurrences of the word "subject" and its variations in different contexts

No Comments

No Known Activations