INDEX

Explanations

words related to power, control, authority, and influence

oai_token-act-pair · gpt-3.5-turbo

terms related to power and control dynamics

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

iate

-0.70

ODE

-0.69

individual

-0.68

ãĥ£

-0.68

rine

-0.67

ECD

-0.66

rum

-0.66

 Scientist

-0.66

ribe

-0.65

aired

-0.65

POSITIVE LOGITS

xual

0.86

anship

0.84

 stemming

0.81

 reversal

0.77

 compulsion

0.77

 prejudice

0.76

 escalation

0.74

uality

0.73

 avoidance

0.73

eering

0.71

Activations Density 0.078%

words related to power, control, authority, and influence

terms related to power and control dynamics

No Comments

No Known Activations