INDEX

Explanations

names of individuals or entities

oai_token-act-pair · gpt-3.5-turbo

references to power dynamics and personal relationships within a competitive or adversarial context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

specified

-0.87

 Regarding

-0.85

ilater

-0.82

artments

-0.81

available

-0.80

 GOODMAN

-0.80

Guest

-0.79

AE

-0.76

Update

-0.76

translation

-0.74

POSITIVE LOGITS

 mentor

1.14

 reputation

1.12

 charisma

1.11

 charismatic

1.10

 legacy

1.09

 arrogance

1.07

 visionary

1.06

 ideals

1.06

ego

1.05

 fortunes

1.05

Activations Density 0.470%

names of individuals or entities

references to power dynamics and personal relationships within a competitive or adversarial context

No Comments

No Known Activations