INDEX

Explanations

the abbreviation or acronym "IL"

oai_token-act-pair · gpt-3.5-turbo

instances of the abbreviation "IL" in various contexts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

uyomi

-0.97

©¶æ

-0.87

perature

-0.81

 Crunch

-0.76

oÄŁ

-0.74

rano

-0.68

 tremend

-0.68

ngth

-0.68

 ende

-0.68

hra

-0.67

POSITIVE LOGITS

IL

1.32

UTION

1.23

IER

1.05

IK

1.00

ibrary

0.98

OAD

0.98

ILS

0.97

ADA

0.95

DER

0.95

UGE

0.93

Activations Density 0.004%

the abbreviation or acronym "IL"

instances of the abbreviation "IL" in various contexts

No Comments

No Known Activations