INDEX

Explanations

mentions of deep emotions or profound issues

oai_token-act-pair · gpt-3.5-turbo

references to deep emotional states or issues

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

pload

-0.79

iannopoulos

-0.75

perty

-0.74

ifully

-0.72

annon

-0.71

Zup

-0.67

theless

-0.66

itely

-0.65

jee

-0.64

 wildly

-0.63

POSITIVE LOGITS

 rooted

1.05

ened

1.00

 pockets

0.96

 seeded

0.93

 roots

0.90

 dives

0.86

 dive

0.86

 seated

0.85

 ingrained

0.83

 vein

0.83

Activations Density 0.050%

mentions of deep emotions or profound issues

references to deep emotional states or issues

No Comments

No Known Activations