INDEX

Explanations

locations and time references

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "in."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ozo

-0.70

 annot

-0.66

 advert

-0.65

 explicitly

-0.61

 synonymous

-0.60

 solic

-0.59

 introductory

-0.59

 autobi

-0.58

 formatted

-0.58

 opposing

-0.58

POSITIVE LOGITS

 Provided

0.87

 Results

0.82

ander

0.81

anders

0.79

tested

0.73

EStream

0.72

results

0.71

Results

0.71

 Nearly

0.71

Redditor

0.70

Activations Density 0.000%

locations and time references

occurrences of the word "in."

No Comments

No Known Activations