INDEX

Explanations

dates and historical events mentioned in a short document

oai_token-act-pair · gpt-3.5-turbo

dates and temporal references

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

ahu

-0.87

entimes

-0.80

 antioxid

-0.77

blogs

-0.77

 chew

-0.70

ĵĺ

-0.70

actions

-0.70

ople

-0.69

parency

-0.66

 providers

-0.65

POSITIVE LOGITS

 onwards

0.88

ruary

0.74

0.73

0.71

0.70

 onward

0.70

 Lyons

0.70

 Kurt

0.69

0.68

Activations Density 0.325%

dates and historical events mentioned in a short document

dates and temporal references

No Comments

No Known Activations