INDEX

Explanations

phrases related to harmful effects or damage

oai_token-act-pair · gpt-3.5-turbo

references to significant negative consequences or damages caused by various factors

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

emouth

-0.76

nesday

-0.75

chens

-0.74

 Ivanka

-0.73

 sugg

-0.73

ceptive

-0.71

 indo

-0.70

 Puzzles

-0.69

 Provided

-0.69

 Allows

-0.68

POSITIVE LOGITS

 havoc

1.12

 severe

0.91

 contamination

0.88

 disturbance

0.87

 premature

0.84

 deterioration

0.83

 damage

0.82

 congestion

0.81

 injury

0.80

 typh

0.80

Activations Density 0.257%

phrases related to harmful effects or damage

references to significant negative consequences or damages caused by various factors

No Comments

No Known Activations