INDEX

Explanations

words related to invasion or aggressive actions

oai_token-act-pair · gpt-3.5-turbo

terms related to invasions and conditions of persuasion

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

lights

-0.74

NAD

-0.69

 unification

-0.67

unal

-0.66

dow

-0.65

 stake

-0.64

 originals

-0.64

 Stru

-0.61

Rud

-0.60

 continuation

-0.60

POSITIVE LOGITS

asions

3.57

asion

2.73

asive

2.60

 invade

1.47

âĦ¢:

1.46

 invaded

1.41

 invading

1.02

ioned

0.96

assion

0.96

ï

0.88

Activations Density 0.040%

words related to invasion or aggressive actions

terms related to invasions and conditions of persuasion

No Comments

No Known Activations