INDEX

Explanations

words related to negative sentiments or emotions

oai_token-act-pair · gpt-3.5-turbo

words related to drug-related themes and the concept of reproach or reclaiming

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

inki

-0.67

 sophistication

-0.65

worms

-0.63

 Vaughan

-0.60

 pockets

-0.60

clerosis

-0.60

ibaba

-0.59

Hem

-0.59

 mercury

-0.59

 narrowing

-0.58

POSITIVE LOGITS

lishes

0.87

nant

0.84

ciation

0.83

lishing

0.82

lisher

0.80

posing

0.78

itory

0.77

ceived

0.75

ciating

0.74

cipled

0.72

Activations Density 0.066%

words related to negative sentiments or emotions

words related to drug-related themes and the concept of reproach or reclaiming

No Comments

No Known Activations