INDEX

Explanations

phrases related to news articles, potentially highlighting specific locations and dates

oai_token-act-pair · gpt-3.5-turbo

occurrences of the acronym "UP."

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

ãĥĸ

-0.70

ãĥł

-0.68

ãĥīãĥ©ãĤ´ãĥ³

-0.67

ãĥĪ

-0.67

Â¯Â¯Â¯Â¯

-0.66

azing

-0.65

obyl

-0.63

////////////////////////////////

-0.62

ãĥĥãĤ¯

-0.62

piece

-0.61

POSITIVE LOGITS

PER

1.11

olicy

1.06

PDATE

1.04

IED

1.00

ERT

0.98

NER

0.93

ERN

0.92

ON

0.91

FER

0.90

rompt

0.90

Activations Density 0.006%

phrases related to news articles, potentially highlighting specific locations and dates

occurrences of the acronym "UP."

No Comments

No Known Activations