INDEX

Explanations

information attributed to a spokesperson

oai_token-act-pair · gpt-3.5-turbo

references to spokespersons in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

dream

-0.69

cffff

-0.67

bows

-0.66

lif

-0.65

NetMessage

-0.65

avery

-0.64

burning

-0.64

panel

-0.62

spir

-0.62

levels

-0.61

POSITIVE LOGITS

 spokesperson

0.79

uted

0.78

 Steph

0.78

atography

0.76

onse

0.76

olicy

0.74

 clarified

0.73

hips

0.73

 emailed

0.72

ials

0.72

Activations Density 0.029%

information attributed to a spokesperson

references to spokespersons in various contexts

No Comments

No Known Activations