INDEX

Explanations

words related to disruptive or destructive actions or events

oai_token-act-pair · gpt-3.5-turbo

terms related to harmful or detrimental actions and influences

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

igon

-0.84

buck

-0.78

elf

-0.76

puter

-0.76

olina

-0.75

hung

-0.75

aintain

-0.73

obyl

-0.72

obbies

-0.72

ille

-0.71

POSITIVE LOGITS

 impulses

0.93

 impulse

0.91

 measures

0.89

 tendencies

0.88

 behav

0.87

 qualities

0.82

 behaviour

0.80

 activity

0.77

 intent

0.76

 gestures

0.74

Activations Density 0.066%

words related to disruptive or destructive actions or events

terms related to harmful or detrimental actions and influences

No Comments

No Known Activations