INDEX

Explanations

approximate numerical values

oai_token-act-pair · gpt-3.5-turbo

terms related to approximation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

Reviewer

-0.81

tab

-0.74

gie

-0.68

keeper

-0.66

ality

-0.65

 Cummings

-0.65

ATA

-0.64

GD

-0.63

SHIP

-0.63

 Courier

-0.62

POSITIVE LOGITS

imates

1.32

rox

1.25

imate

1.22

imating

1.16

imation

1.04

imated

1.02

 approx

0.98

 approximate

0.92

imately

0.88

actly

0.87

Activations Density 0.009%

approximate numerical values

terms related to approximation

No Comments

No Known Activations