INDEX

Explanations

phrases that express reasons or causes

oai_token-act-pair · gpt-3.5-turbo

instances of the phrase "for" related to various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

Enlarge

-0.65

Norm

-0.62

Introduced

-0.62

Witness

-0.61

Force

-0.59

cles

-0.59

 Norn

-0.58

params

-0.58

DIT

-0.58

minimum

-0.58

POSITIVE LOGITS

bidden

1.25

geries

1.11

gotten

1.04

gery

1.00

 example

0.97

 instance

0.95

ked

0.94

aging

0.94

 sale

0.91

bid

0.90

Activations Density 0.234%

phrases that express reasons or causes

instances of the phrase "for" related to various contexts

No Comments

No Known Activations