INDEX

Explanations

instances where someone is able to perceive or deduce information

oai_token-act-pair · gpt-3.5-turbo

instances of the word "tell" in various contexts

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

urdue

-0.85

cdn

-0.78

ilings

-0.72

Merit

-0.70

pmwiki

-0.70

ItemTracker

-0.69

egal

-0.69

tu

-0.63

Naz

-0.63

mitted

-0.63

POSITIVE LOGITS

tale

1.25

ingly

0.99

tell

0.83

 anecd

0.73

ariat

0.71

»

0.70

ively

0.68

Ĭ±

0.67

 lies

0.65

 Tell

0.64

Activations Density 0.019%

instances where someone is able to perceive or deduce information

instances of the word "tell" in various contexts

No Comments

No Known Activations