INDEX

Explanations

phrases related to criticism or negative sentiment

oai_token-act-pair · gpt-3.5-turbo

negative sentiments about specific groups or concepts

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

SEA

-0.67

iterranean

-0.66

miah

-0.64

SPONSORED

-0.60

©¶æ¥µ

-0.60

ONSORED

-0.59

abama

-0.59

theless

-0.58

pherd

-0.58

arthed

-0.57

POSITIVE LOGITS

 needles

0.63

shit

0.62

bro

0.61

 yarn

0.60

Redditor

0.59

 saline

0.59

cs

0.59

 pops

0.59

ho

0.58

bots

0.58

Activations Density 2.275%

phrases related to criticism or negative sentiment

negative sentiments about specific groups or concepts

No Comments

No Known Activations