INDEX

Explanations

instances where individuals are vocalizing opinions or criticisms

oai_token-act-pair · gpt-3.5-turbo

instances of speaking out or expressing opinions against various issues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 rede

-0.64

vous

-0.60

 Gadget

-0.60

liction

-0.59

 rolls

-0.59

ahon

-0.58

 turnover

-0.58

 Lann

-0.58

 oxidation

-0.57

oreal

-0.57

POSITIVE LOGITS

stretched

0.99

 loud

0.94

 louder

0.82

 loudly

0.81

mbuds

0.78

lier

0.78

eka

0.76

rage

0.76

olate

0.75

burst

0.72

Activations Density 0.029%

instances where individuals are vocalizing opinions or criticisms

instances of speaking out or expressing opinions against various issues

No Comments

No Known Activations