INDEX

Explanations

words related to traveling to specific locations

oai_token-act-pair · gpt-3.5-turbo

instances of the word "to," indicating a focus on travel or movement context

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

 complied

-0.72

spir

-0.67

 compounded

-0.66

standing

-0.65

pie

-0.64

 headed

-0.63

 dealt

-0.63

 Cust

-0.62

 phased

-0.62

 handled

-0.61

POSITIVE LOGITS

 Disneyland

0.86

pload

0.84

 Vegas

0.82

pless

0.79

 meet

0.78

 Omaha

0.77

 fulfill

0.77

 Liberia

0.77

 relieve

0.77

 Columbus

0.75

Activations Density 0.193%

words related to traveling to specific locations

instances of the word "to," indicating a focus on travel or movement context

No Comments

No Known Activations