INDEX

Explanations

explicitly stated or specified information

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "explicit" in various contexts related to clarity and directness

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

 Tycoon

-0.91

STON

-0.81

Kenn

-0.72

«ĺ

-0.72

Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯

-0.72

Park

-0.72

pered

-0.71

Score

-0.70

Squ

-0.70

ADS

-0.69

POSITIVE LOGITS

 guiActiveUn

0.93

 explicit

0.85

 textual

0.84

 explicitly

0.77

bidden

0.73

ities

0.71

 prescribing

0.71

 hint

0.71

 contractual

0.70

 explor

0.70

Activations Density 0.025%

explicitly stated or specified information

occurrences of the word "explicit" in various contexts related to clarity and directness

No Comments

No Known Activations