INDEX

Explanations

terms related to fortifications or strongly held positions

oai_token-act-pair · gpt-3.5-turbo

mentions of fortifications or related defensive structures

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

çīĪ

-0.73

hner

-0.73

 Spectrum

-0.72

VIDEO

-0.72

lihood

-0.71

EV

-0.70

ĪĴ

-0.69

GROUP

-0.69

VID

-0.69

 Results

-0.66

POSITIVE LOGITS

uitous

1.60

ifications

1.25

ieth

1.15

ification

1.10

resses

1.09

 fort

1.06

ifying

1.02

itude

0.98

itious

0.94

ified

0.94

Activations Density 0.024%

terms related to fortifications or strongly held positions

mentions of fortifications or related defensive structures

No Comments

No Known Activations