INDEX

Explanations

references to time, particularly phrases indicating the passage of time

oai_token-act-pair · gpt-3.5-turbo

phrases related to time elapsed

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

pert

-0.82

ilet

-0.74

ulkan

-0.72

ulhu

-0.70

undrum

-0.67

thinkable

-0.65

 Offic

-0.64

 Overs

-0.64

cients

-0.63

acus

-0.63

POSITIVE LOGITS

aneously

0.82

 enough

0.78

nor

0.78

Enough

0.72

ago

0.71

 Enough

0.70

 Pradesh

0.69

 suffice

0.68

terday

0.67

letters

0.66

Activations Density 0.031%

references to time, particularly phrases indicating the passage of time

phrases related to time elapsed

No Comments

No Known Activations