INDEX

Explanations

mentions of the word "tipped" in various contexts

oai_token-act-pair · gpt-3.5-turbo

instances of the word "tipped" and variations concerning tipping points and actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Address

-0.68

 isol

-0.67

mem

-0.66

 Memory

-0.64

 abstract

-0.63

 Instruction

-0.62

statement

-0.62

iem

-0.61

 memor

-0.61

 solitude

-0.60

POSITIVE LOGITS

 tipped

4.25

 tipping

2.40

tip

1.53

 tips

1.29

tip

1.13

 tilted

1.12

 alerted

1.08

 pegged

1.04

Tip

0.99

 toppled

0.96

Activations Density 0.016%

mentions of the word "tipped" in various contexts

instances of the word "tipped" and variations concerning tipping points and actions

No Comments

No Known Activations